Robot AI model
OpenVLA
OpenVLA adapts a Prismatic vision-language backbone to tokenized robot actions; its authors describe the policy as 7B. The code and checkpoint card specify MIT, while the repository explicitly states that pretrained models also inherit the Llama 2 Community License Agreement.
Records last verified Oct 13, 2026 · Send a correction
Also known as OpenVLA-7B
- Organization
- OpenVLA Collaboration
- Type
- VLA
- Released / introduced
- Jun 13, 2024
- Public weights
- Public weights
- Licence / scope
- MIT; Llama 2 Community License Agreement
- Parameters
- 7B
- Base model
- Prismatic prism-dinosiglip-224px (Llama 2, DINOv2 and SigLIP)
- Humanoid association
- None reported
- Inputs
- vision, language
- Outputs
- end-effector action tokens
Release dates can mark an announcement or paper introduction. Public checkpoints, code licences and weight terms are distinct; follow the publisher’s links for the stated scope.
Documented embodiments
Source-associated hardware; this list does not establish plug-and-play compatibility or commercial robot capability.
- WidowX
The model card documents BridgeData V2 zero-shot control.
Disclosed training data
More robot AI models
Evidence
Sources
- [s1] OpenVLA: An Open-Source Vision-Language-Action Model ↗
OpenVLA Collaboration · paper · published Jun 13, 2024 · accessed Oct 13, 2026
- [s2] OpenVLA-7B official model card ↗
OpenVLA Collaboration · official · accessed Oct 13, 2026
- [s3] OpenVLA official repository: pretrained models and inherited license restrictions ↗
OpenVLA Collaboration · official · accessed Oct 13, 2026
- [s4] Llama 2 Community License Agreement ↗
Meta · official · accessed Oct 13, 2026